cs229fall2017

Stanford CS229 Fall 2017 - Team Project - Real Time Tennis Match Prediction Using Machine Learning

Team: Eddie Chen (yc4@), Yubo Tian (yubotian@), Yi Zhong (yizhon@)

Project Idea Fair

Data generation:

• For player model in current match, set G_USE_DIFF in .\data_scripts\curr_match_data.py to False, run final_join.py (default)
• For difference model in current match, set G_USE_DIFF in .\data_scripts\curr_match_data.py to True, run final_join_diff.py

Train model, and plot learning curves:

• After obtaining final_joined_data (default) or final_joined_data_diff, run model_evaluation.py for model analysis
• After obtaining final_joined_data (default) or final_joined_data_diff, run graph.py for graphs

File description:

• logistic.py, svm.py: contains functions related to logistic and svm
• model_selection.py: contains functions related to selecting models with rfe
• hist_match_pred.py: initial modeling using historical data only
• curr_match_pred.py: initial modeling using current data only
• joined_data_pred.py: initial modeling using historical + current data